cancel
Turn on suggestions
Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.
Showing results for 
Search instead for 
Did you mean: 
Veritas.com
Support

Best Practice Question

Erik_Pierson
Level 2

‎11-12-2007 03:18 PM

I would like suggestions on setting up BackupExec with the following scenario:
 
  • All backup jobs will run to disk (VTL)
  • Once the backup jobs complete, they will automatically duplicate to tape (tape library)
  • Tapes will then go off site each day.
  • Friday full backups
  • Monday - Thursday incremental backup.
 
I have about 2.5 TB of data to back up, and about 20 servers.  All are Windows Servers. I also have Exchange 2003 and SQL 2000/2005 servers to backup.  The Exchange IS is pretty large, around 200GB.
 
Do you think I should run multiple jobs each night (for example, one job for Exchange, one for SQL, and one for file servers, etc) or should I just lump them all together?  If I split them up, does it make sense to stagger them so once the VTL job finishes, the duplicate job does not hit the tape all at the same time?
 
Also, I am planning on using the Grandfather, Father, Son tape rotation.  I will keep tapes off-site for one month and month-end tapes off-site for three years. 
 
Let me know if you have any thoughts!
 
0 Kudos
1 REPLY 1

Hywel_Mallett
Level 6
Certified

‎11-13-2007 04:07 PM

Do you think I should run multiple jobs each night (for example, one job for Exchange, one for SQL, and one for file servers, etc) or should I just lump them all together?

I would definitely split them up. It would work if you had them all together, but splitting them allows you to have different options for each backup type, and apply different policies to each.

If I split them up, does it make sense to stagger them so once the VTL job finishes, the duplicate job does not hit the tape all at the same time?

It depends. If the VTL can cope with multiple jobs without performance degrading, then let them run at the same time, and the duplicate jobs should be queued, so they run one after another.
0 Kudos